It is well known that most of the bodily states in mammals including most disease states, are affected by proteins. Such proteins, either acting directly or through their enzymatic functions, contribute in major proportion to many diseases in animals and man. Classical therapeutics has generally focused on interactions with such proteins in efforts to moderate their disease causing or disease potentiating functions. Recently, however, attempts have been made to moderate the actual production of such proteins by interactions with molecules that direct their synthesis, such as intracellular RNA. By interfering with the production of proteins, it has been hoped to affect therapeutic results with maximum effect and minimal side effects. It is the general object of such therapeutic approaches to interfere with or otherwise modulate gene expression leading to undesired protein formation.
One method for inhibiting specific gene expression is the use of oligonucleotides and oligonucleotide analogs as "antisense" agents. The oligonucleotides or oligonucleotide analogs complimentary to a specific, target, messenger RNA (mRNA) sequence are used. Antisense methodology is often directed to the complementary hybridization of relatively short oligonucleotides and oligonucleotide analogs to single-stranded mRNA or single-stranded DNA such that the normal, essential functions of these intracellular nucleic acids are disrupted. Hybridization is the sequence specific hydrogen bonding of oligonucleotides or oligonucleotide analogs to Watson-Crick base pairs of RNA or single-stranded DNA. Such base pairs are said to be complementary to one another. The oligonucleotides and oligonucleotide analogs are intended to inhibit the activity of the selected mRNA--to interfere with translation reactions by which proteins coded by the mRNA are produced--by any of a number of mechanisms. The inhibition of the formation of the specific proteins that are coded for by the mRNA sequences interfered with have been hoped to lead to therapeutic benefits. Cook, P. D. Anti-Cancer Drug Design 1991, 6, 585; Cook, P. D. Medicinal Chemistry Strategies for Antisense Research, in Antisense Research & Applications, Crooke, et al., CRC Press, Inc.; Boca Raton, Fla., 1993; Uhlmann, et al., A. Chem. Rev. 1990, 90,543.
Oligonucleotides and oligonucleotide analogs are now accepted as therapeutic agents holding great promise for therapeutics and diagnostics methods. But applications of oligonucleotides and oligonucleotide analogs as antisense agents for therapeutic purposes, diagnostic purposes, and research reagents often require that the oligonucleotides or oligonucleotide analogs be synthesized in large quantities, be transported across cell membranes or taken up by cells, appropriately hybridize to targeted RNA or DNA, and subsequently terminate or disrupt nucleic acid function. These critical functions depend on the initial stability of oligonucleotides and oligonucleotide analogs toward nuclease degradation. A serious deficiency of unmodified oligonucleotides for these purposes, particularly antisense therapeutics, is the enzymatic degradation of the administered oligonucleotides by a variety of intracellular and extracellular ubiquitous nucleolytic enzymes, hereinafter referred to as "nucleases."
Initially, only two mechanisms or terminating events have been thought to be operating in the antisense approach to therapeutics. These are the hybridization arrest mechanism and the cleavage of hybridized RNA by the cellular enzyme, ribonuclease H (RNase H). Cook, 1991; supra; Cook, 1993, supra; Uhlmann, supra; Walder, et al., Proc. Natl. Acad. Sci., USA, 1988, 85, 5011; Dagle, et al., Antisense Research & Development, 1991, 1, 11. It is likely, however, that additional "natural" events may be involved in the disruption of targeted RNA. Many of these naturally occurring events are discussed in Oligonucleotides: Antisense Inhibitors of Gene Expression, CRC Press, Inc., Boca Raton, Fla. (Cohen ed., 1989).
Hybridization-arrest denotes the terminating event in which the oligonucleotide inhibitor binds to the target nucleic acid and thus prevents, by simple steric hindrance, the binding of essential proteins, most often ribosomes, to the nucleic acid. Methyl phosphonate oligonucleotides, Miller, et al., Anti-Cancer Drug Design, 1987, 2, 117-128, and .alpha.-anomer oligonucleotides are two extensively studied antisense agents that are thought to disrupt nucleic acid function by hybridization arrest.
The second "natural" type of terminating event is the activation of RNase H by the heteroduplex formed between the DNA type oligonucleotides or oligonucleotide analogs and the targeted RNA with subsequent cleavage of target RNA by the enzyme. The oligonucleotides or oligonucleotide analogs, which must be of the deoxyribose type, hybridize with the targeted RNA and this duplex activates the RNase H enzyme to cleave the RNA strand, thus destroying the normal function of the RNA. Phosphorothioate modified oligonucleotides are the most prominent example of antisense agents that are thought to operate by this type of antisense terminating event. Walder, supra and Stein, et al., Nucleic Acids Research, 1988, 16, 3209-3221 describe the role that RNase H plays in the antisense approach.
A number of chemical modifications have been introduced into antisense agents--oligonucleotides and oligonucleotide analogs--to increase their therapeutic activity. Such modifications are designed to increase cell penetration of the antisense agents, to stabilize the antisense agents from nucleases and other enzymes that degrade or interfere with their structure or activity in the body, to enhance the antisense agents' binding to targeted RNA, and to provide a mode of disruption (terminating event) once the antisense agents are sequence-specifically bound to targeted RNA, and to improve the antisense agents' pharmacokinetic and pharmacodynamic properties. It is unlikely that unmodified, "wild type," oligonucleotides will be useful therapeutic agents because they are rapidly degraded by nucleases. A primary focus of antisense research has been to modify oligonucleotides to render them resistant to such nucleases. These modifications are designed to enhance the uptake of antisense agents--oligonucleotides and oligonucleotide analogs--and thus provide effective therapeutic, research reagent, or diagnostic uses.
To increase the potency via the "natural" termination events, the most often used oligonucleotide modification is modification at the sugar-phosphate backbone, particularly on the phosphorus atom. Phosphorothioates, methyl phosphonates, phosphoramidites, and phosphorotriesters have been reported to have various levels of resistance to nucleases. Backbone modifications are disclosed as set forth in U.S. patent applications assigned to a common assignee hereof, entitled "Backbone Modified Oligonucleotide Analogs," Ser. No. 703,619 and "Heteroatomic Oligonucleotide Linkages," Ser. No. 903,160, the disclosures of which are incorporated herein by reference to disclose more fully such modifications.
An example of phosphate modifications include methyl phosphonate oligonucleotides, where the phosphoryl oxygen of the phosphorodiester linking moiety is replaced with methylene groups or the nucleotide elements together are replaced, either in total or in part, by methyl groups. Other types of modifications to the phosphorus atom of the phosphate backbone of oligonucleotides include phosphorothioate oligonucleotides. The phosphorothioate modified oligodeoxynucleotides are capable of terminating RNA by activation of RNase H upon hybridization to RNA although hybridization arrest of RNA function may play some part in their activity. Phosphoramidites have been disclosed as set forth in U.S. patent application assigned to a common assignee hereof, entitled "Improved Process for Preparation of 2'-O-Alkylguanoslnes and Related Compounds," Ser. No. 918,362, now U.S. Pat. No. 5,506,351 the disclosures of which are incorporated herein by reference to disclose more fully such modifications. However, all reported modifications of the sugar-phosphate backbone, with the exception of phosphorothioates and phosphorodithioates, obliterate the RNase H terminating event. Cook, 1991, supra; Cook, 1993, supra; Uhlmann, supra. Heteroduplexes formed between RNA and oligodeoxynucleotides bearing 2'-sugar modifications, RNA mimics such as fluoro and alkoxys, do not support RNase H-mediated cleavage. These modified heteroduplexes assume an A form helical geometry as does RNA-RNA heteroduplexes which also do not support RNase H cleavage. Kawasaki, et al., J. Med. Chem., in press 1993; Lesnik, et al., Biochemistry, submitted 1993; Inoue, et al., Nucleic Acids Res., 1987, 15, 6131.
Other modifications to "wild type" oligonucleotides made to enhance resistance to nucleases, activate the RNase terminating event, or enhance the RNA-oligonucleotide duplex's hybridization properties include functionalizing the nucleoside's naturally occurring sugar. Sugar modifications are disclosed as set forth in PCT Application assigned to a common assignee hereof, entitled "Compositions and Methods for Detecting and Modulating RNA Activity and Gene Expression," PCT Patent Application No. PCT/US 91/00243, International Publication No. WO 91/10671, the disclosures of which are incorporated herein by reference to disclose more fully such modifications.
Other synthetic terminating events, as compared to hybridization arrest and RNase H cleavage, have been studied in an attempt to increase the potency of oligonucleotides and oligonucleotide analogs for use in antisense diagnostics and therapeutics. One area of research is based on the concept that antisense oligonucleotides with modified heterocyclic portions, rather than sugar-phosphate modifications, can be resistant to nucleolytic degradation, yet on hybridization to target RNA provide a heteroduplex that supports RNase H-mediated cleavage. Modifications in the heterocycle portion of oligonucleotides may not affect the heteroduplex helical geometry of sugar that is necessary for RNase H cleavage.
Another approach is directed to the development of sequence-specific chemical RNA cleavers. This concept requires attaching pendent groups with acid/base properties to oligonucleotides. The pendent group is not involved with the specific Watson-Crick hybridization of the oligonucleotides or oligonucleotide analogs with mRNA but is carried along by the oligonucleotide or oligonucleotide analog to serve as a reactive functionality. The pendent group is intended to interact with mRNA in some manner more effectively to inhibit translation of mRNA into protein. Such pendent groups have also been attached to molecules targeted to either single or double stranded DNA. Such pendent groups include, intercalating agents, cross-linkers, alkylating agents, or coordination complexes containing a metal ion with associated ligands.
The sites of attachment of the pendent groups to oligonucleotides and oligonucleotide analogs play an important, yet imperfectly known, part in the effectiveness of oligonucleotides and oligonucleotide analogs for therapeutics and diagnostics.
The half life of the formed RNA-oligonucleotide or oligonucleotide analog duplex may be greatly affected by the positioning of the tethered functional group containing the reactive functionality. Inappropriate positioning of reactive functional groups, such as placement on the Watson-Crick base pair sites, would likely preclude duplex formation. Other attachment sites may potentially allow sequence-specific binding but may be of such low stability that the reactive functionality will not have sufficient time to initiate RNA disruption.
A stable RNA-oligonucleotide or oligonucleotide analog heteroduplex is believed to be important, because without a sufficient half life the reactive or non-reactive functionalities of this invention may not have enough time to initiate the cleavage or otherwise disrupt RNA function. Improved complementation between modified oligonucleotides or oligonucleotides and targeted RNA will likely result in the most stable heteroduplexes.
Targeted RNA is inactivated by formation of covalent links between a modified oligonucleotide and the RNA 2'-hydroxyl group. A variety of structural studies such as X-ray diffraction, chemical reaction, and molecular modeling studies suggests that the 2'-hydroxyl group of RNA in a duplex or heteroduplex resides in the minor groove. The minor side or minor groove of the duplexes formed between such oligonucleotides or modified oligonucleotides and the targeted RNA has been found to be the greatly preferred site for functional group activity.
Prior approaches using cross-linking agents, alkylating agents, and radical generating species as pendent groups on oligonucleotides for antisense diagnostics and therapeutics have had several significant shortcomings. Prior workers have described most pendent groups as being attached to a phosphorus atom which affords oligonucleotides and oligonucleotide analogs with inferior hybridization properties. A phosphorus atom attachment site can allow a reactive group access to both the major and minor grooves. However, internal phosphorus modification results in greatly reduced heteroduplex stability. Attachments at the 3' and/or 5' ends are limiting in that only one or two functional groups can be accommodated in the oligonucleotide compositions.
Other approaches have included attaching reactive functionalities or pendent groups to the 5-position of thymine, and the 7-position of purines. Functionalities placed in the 5-position or 7-position of bases, pyrimidine and purine, respectively, will typically reside in the major groove of the duplex and will not be in proximity to the RNA 2'-hydroxyl substrate. The 2'-hydroxyl is a "trigger" point for RNA inactivation, and thus, any reactive functionalities should be in appropriate proximity to the receptive substrate located in the targeted RNA, especially the most sensitive point, the 2'-hydroxyl group.
Some workers have looked at substitutions at the N-2 position of certain purines, such as hypoxanthine, guanine or adenine. See, e.g., Harris et al., J. Am. Chem. Soc'y, 1991, 113, 4328-4329; Johnson et al., J. Am. Chem. Soc'y, 1992, 114, 4923-4924; Lee et al., Tetrahedron Letters, 1990, 31, 6773-6776; Casale et al.. J. Am. Chem. Soc'y, 1990, 112, 5264-5271.
The functionalities' point of attachment to the base units, which in turn may be converted to modified oligonucleotides, might be considered important in the design of compositions for sequence-specific destruction or modulation of targeted RNA. It is important that the functionalities not interfere with Watson-Crick base pair hydrogen bonding rules, as this is the sequence-specific recognition/binding factor essential for selection of the desired RNA to be disrupted. Further, the functionalities preferably should improve the oligonucleotides compositions' pharmacokinetic and/or pharmacodynamic properties, as well as the oligonucleotide compositions' transport properties across cellular membranes. It is also important that the pendent groups designed to support either enzymatic or chemical cleavage of RNA must be compatible with the requisite hybridization step. When hybridized to RNA, the pendent groups would be accessible, via the minor groove, to the 2'-hydroxyl and phosphorodiester linkages of the targeted RNA.
These aforementioned prior attempts have been relatively insensitive, that is the reactive pendent groups have not been effectively delivered to sites on mRNA molecules for alkylation or cleavage in an effective proportion. Moreover, even if the reactivity of such materials were perfect, (i.e., if each reactive functionality were to actually react with a mRNA molecule), the effect would be no better than stoichiometric. That is, only one mRNA molecule would be inactivated for each oligonucleotide or oligonucleotide analog molecule. It is also likely that the non-specific interactions of oligonucleotide compositions with molecules other then the target RNA, for example with other molecules that may be alkylated or which may react with radical species, as well as self-destruction, not only diminishes the diagnostic or therapeutic effect of the antisense treatment but also leads to undesired toxic reactions in the cell or in vitro. This is especially acute with the radical species that are believed to be able to diffuse beyond the locus of the specific hybridization to cause undesired damage to non-target materials, other cellular molecules, and cellular metabolites. This perceived lack of specificity and stoichiometric limit to the efficacy of such prior alkylating agents and radical generating-types of antisense oligonucleotide compositions is a significant drawback to their employment.
Accordingly, there remains a great need for antisense oligonucleotide compositions that are capable of improved specificity and effectiveness both in binding and modulating mRNA modulation or inactivating mRNA without imposing undesirable side effects. The present invention addresses these, as well as other, needs by presenting novel compounds, based on the purine ring system, that may be used as oligonucleotide intermediates. It has now been found that certain positions on the nucleosides of double stranded nucleic acids are exposed in the minor groove and may be substituted without affecting Watson-Crick base-pairing or duplex stability. Reactive or non-reactive functionalities placed in these positions can best initiate cleavage and destruction of targeted RNA or interfere with its activity.